Re: [unrev-II] How DKR Penetration Will Be Achieved

From: Peter P. Yim (
Date: Sun Feb 27 2000 - 19:29:32 PST

From: "Peter P. Yim" <>


What an opportune time, against the backdrop of your discussion here,
for me to introduce (to those unfamiliar with the work) the
"Language/Action Perspective" in system design. This is seminal work by
Fernando Flores and Terry Winograd which formed the basis for groupware
products like "the Coordinator", various "Workflow" products (e.g.
Action Technologies' "Metro <>), as well as
"CIM3" (Computer Integrated Man-Machine Manufacturing), a framework
which I had started developing since the mid eighties.

Our (Winograd, Newman & Yim) earlier paper "Including People in Computer
Integrated Manufacturing Designs" (at
<>) can be a convenient starting
point. You may follow the references and links from there if you find
the approach interesting. We argue that design approaches that assume
the "ideal" (e.g. assuming that things are fully describable) would fall
short in its ability to model reality.

I believe Doug Engelbart pushed "augmentation" rather than "full
automation" probably for the same reasons. I would interpret his
insistence on "clear, unambiguous, definitions of nouns and verbs" as
being part of the pursuit towards higher performance, rather than being
a prerequisite for CoDIAK to function properly.


P.S. I promised Dick Karpinsky the link to this paper last Thursday. So
here it is, Dick (sorry for giving you the wrong URL at class, Dick).


Henry van Eyken wrote Sun, 27 Feb 2000 19:21:23 -0500 :
> Eric Armstrong wrote Sun, 27 Feb 2000 14:47:37 -0800:
> > One of the thornier issues to solve is identifying situations where
> > miscommunication occurs because different people have different meanings
> > for the same term -- or the same meaning for different terms. It will
> > require extremely high levels of intelligence to identify and resolve
> > such problems. That may be two centuries out, if it is solvable at all.
> Indeed. D.E. insists on words (nouns, verbs) having unambiguous meanings, but
> there is so much meaning in phrasing as well.
> I understand that English has about 30 percent redundancy for allowing
> listeners to zero in on what speakers are trying to transmit. Lawyers try to
> get around this by sense-dulling refinement. Math and computer languages cut
> the chase to zero percent redundancy. Natural language interpreters try to
> teach computers to interpret natural language. But natural languages change
> continually.
> Might it be possible to "discipline" a natural language so as to make
> communication with computers free from misunderstanding? I imagine some people
> must be working on that somewhere ...
> Henry

Why wait for a credit decision? Get one instantly! Apply for
the Capital One 9.9% Fixed APR Visa Platinum!

Community email addresses:
  Post message:
  List owner:

Shortcut URL to this page:

This archive was generated by hypermail 2.0.0 : Tue Aug 21 2001 - 18:56:53 PDT